Will Record Fallout Make 2.5G a Risky Business?

Comments
Print
Service providers lose 2 percent to 11 percent of their annual revenue, say industry analysts. While fraud makes up a large chunk of the losses (18 percent, according to Cap Gemini Ernst & Young), network elements and billing and operation support systems account for 67 percent of revenue leaks.

These costly losses, estimated in 2000 and 2001 reports, occur within the familiar, predictable wireline environment. What the losses will be in a dynamic, data-centric 2.5G environment are unknown, but everyone’s making a safe bet that they will be considerably higher.

“Revenue leakage will get much worse as data services are introduced,” says Jonjie Sena, director of product management at Ace-Comm. “A wireless phone produces a single CDR. When we move to event-based services, the opportunity for corrupted, invalid, lost and incorrectly recorded data becomes exponentially higher.”

Moving from voice to next-generation data services and adding GPRS and EDGE devices create a billing environment ripe for lost records. Adding unfamiliar applications, nontraditional network elements and out-of-network content servers adds even more leaks.

Even with today’s small number of subscribers using 2.5G services, mobile operatorsare already taking a hit. Video downloads, location-based services, picture messages,MP3 music downloads and even SMS text messages are testing providers and forcingthe development of creative, bulletproof solutions that prevent the loss of datarecords. (See “RA Programs: Locating Loose Change and Hard Cash”, fora discussion of the current spate of revenue assurance programs hitting the telecommunicationsmarket.)

Known Offenders

Everywhere an operator looks—every system, platform and node—is a potential revenue hole. While many leaks occur at the network level, the opportunity for revenue loss can occur at any point in the services supply chain.

Leaks often occur, for example, between the prepaid system and billing system. Amdocs recounts how one European operator let customers send a text message to activate 2.5G services, such as MMS. The operator configured the network correctly to provision the service by SMS, but did not configure the prepaid system to catch the subscriber information and deliver it to the billing system. The subscriber services were activated but never billed. For this specific configuration error, Amdocs estimated losses at $1.5 million over three or four months, from when the operator introduced the new services until the leak was discovered.

MMS, picture messaging and MP3 offerings add QoS parameters into the revenue leakage mix.

Many mediation platforms do not support QoS. Not all systems can correlate and aggregate based on QoS, in particular the type of QoS that 2.5G services require, such as bytes vs. packets or set time parameters.

2.5G services demand QoS. When picture messaging was first introduced, many subscribers were shocked at the poor quality of the pictures. To make the service valuable—and not just a gimmick—operators must tie QoS parameters to the user, session and application.

Portal also worked with its customers to close revenue gaps. When Telus first introduced text messaging, WAP services, games and 411 look-ups, it chose Portal to handle pre- and postpaid rating for data. The location-based services, such as 411 look-ups, were not 100 percent leak-free, due to limited exposure between the Lucent IN, which was handling prepaid voice, and Portal. The Lucent IN was unable to transfer subscriber information in real time to Portal’s Infranet. For these sub-dollar transactions, though, Telus was willing to have a narrow window of leakage. Before Telus rolls out higher-value data services, it intends to add real-time integration.

A major concern with such 2.5G services is the complicated chain of events. A single interrupted activity will disrupt the service and incur lost revenue. Let’s say a subscriber requests a service that sends a text message with preflight information. Flight time and gate location is messaged to the subscriber two hours before departure time. For that service to succeed, both the server for the location-based service and the SMS center must be working. If the messaging part wasn’t working on the radio segment or the short message center couldn’t read the request, the message will be dropped.

Telenor, another Portal customer, experienced revenue loss from its SMS prepaid offering before deploying Infranet. To test demand, Telenor turned up text messaging for its prepaid customers. With absolutely no promotion of the service, prepaid subscribers were sent 22 million messages over three weeks. The unadvertised offering was such a success that it flooded the control center and crashed the systems, requiring Telenor to turn the service down. Now aware of the pent-up demand, Telenor built a homegrown billing system for its prepaid system and turned the service up again. The internal system had a one- to two-hour delay, allowing many prepaid subscribers to continue text messages even after they had spent their balance. Since Telenor deployed Portal software into its prepaid environment, leakage is less than 1 percent, says James Morehead, senior director of solutions marketing for Portal. The remaining leakage is due to non-real-time interfaces still present in the legacy environment.

To prevent these types of losses (as well as whet subscriber appetites for 2.5G services), many operators are offering a period of free service. The IT and marketing staffs use this time to test for interoperability, examine QoS, gather usage statistics and patterns, and define rating plans.

“An introductory, free period is the smart, sensible way to go,” says Sena at Ace-Comm. “Operators can do some testing and try to understand how subscribers are using the service and define patterns. One of the major problems with 2.5G services is that we have no realistic view of patterns. What happens when the service is in use, what is the hit on the network, what is the hit on the nodes, and what demographics are using the service are all unknowns.”

Answering those questions is a concern for operations, but in many cases competition is forcing them to offer services before they can bill for them. “When operators are forced to launch quickly, something has to give—and that something is usually billing or mediation,” says Malcolm Lewis, director of product management for mobile services at CSG. “Billing and mediation get a huge number of requests for change. And usually each new request is queued up behind eight or more other requests. In those cases, time to market takes precedence.”

Swedish operators are offering free MMS, in one interesting example of gratis data services. Comviq, Vodafone and Telia, which launched MMS services on their GPRS networks in late 2002, were still offering the service free in March. The freebie could last until June.

It’s unclear why operators are not charging for MMS, but some insiders have suggested that the billing systems can’t bill per message. Unlike other services that are billed per bit, byte or time, MMS must be billed per message.

The CDR Stops Here

In the wireline environment, the network elements, billing systems and mediation platforms are the main culprits for lost revenue. You can expect these areas will be soft spots in the 2.5G environment, too. “Today’s leakage is not different from leaks 10 years ago,” says Yves Robinson, director of product management at Vibrant Solutions. “Generating the data and event records isn’t the challenge. It’s the overwhelming numbers of incomplete records that fall out of the system.”

Records fall out of the billing system—or never reach the system—for any number of reasons. The CDR may generate an error because it doesn’t contain the right information, or it has incomplete information. Or the data may be correct but in the wrong field, or have the wrong format, or contain incorrect header or footer information. The data could be wrong due to a configuration error on the switch. Records also fall out due to bad reference table values that were not updated correctly to correlate with billing and mediation.

2.5G operators can expect an additional degree of record fallout due to undefined event structures. “The record could come to mediation and if the mediation platform says ‘I don’t know what to do with this,’ it will drop the record,” says Kim Jones, product marketing manager at Convergys. “No one knows what the structures will be in 2.5G, but items could include an IP address, domain name, QoS, time or date stamp. If defined incorrectly, any of these items could prevent the service from being a billable service.”

Another key concern for mobile operators is the application services and other network nodes outside of their operations center control. According to Amdocs’ Guy Alon, revenue assurance product manager, any change from either side will open the network to revenue loss.

2.5G data services require interfaces between the service provider’s network and the content providers. “When one partner adjusts or changes its data structure or even frequency,” he says, “that change in the data layer could result in a dropped file.”

In the past, dropped records fell into a pit and were never recovered. When a CDR could not be transferred to the billing system, no matter what the reason, the record’s data was lost. (Phillips Group forecasts that 12 percent of records wind up in the error file.) Today’s intelligent billers are adding functionality that runs a lost record through various systems to recover and correct the data so that the billing system can process the CDR.

“Next-gen billers have built-in functionality that does more pushing and pulling of data between warehousing systems to capture data and rerun it through the systems,” says Convergys’ Jones. “In legacy environments, error records never get to billing. We can detect errors and run the records back through the rating catalog or billing catalog to look at the event file and see what went wrong.”

These types of intelligent billers are useful in 2.5G environments, but operators must also consider record errors that arise from their content partners. “It’s easy to apply corrective actions for predefined rules, or change the value or format and recycle the record to the originating system, if you are responsible for all the systems,” says Amdocs’ Alon. “When you add out-of-network elements in a next-generation environment, operators will have to create an error mechanism for each of their partners.”

Record fallout isn’t mobile operators’ only concern. They must also have an effective mediation platform that can correlate all the incoming events. In some cases the records may never reach the mediation platform. The media gateway could drop CDRs during transfer, or data could be lost if an element does not have enough storage to accommodate all the incoming records. If the memory is full, any further data is dropped and never passed to mediation. Configuration poses another problem. If the mediation platform is not configured to read the records from the node, the records are dropped.

Even with the chance that some data records could be lost before they reach the mediation platform, mediation systems will still have to handle plenty of data. “The volume of records in a 2.5G environment is estimated to be 15 to 50 times higher than in a traditional mobile world,” says Katri Sahlman, senior vice president for product business at Comptel. “But the volume of data is not a problem. The problem is the correlation of the usage and the services.”

Data services produce more events from multiple elements. At least five different elements produce event records. Ace-Comm’s Sena explains the challenges of aggregating and correlating data sessions using the example of session initiation protocol (SIP). “You have records from the SCP [service creation platform] and the media feed from the server. These records must be correlated to identify total usage. Usage is rarely visible from the end-point records; you also need data from the media gateway and the SIP session.”

When bundled services are added to the mix, the correlation and revenue assurance become even more complex. If a subscriber is notified about videos through SMS, operators face multiple challenges, says Danny Itzigsohn, marketing wireless solution manager at TTI Telecom, especially if the subscriber is moving between cells. “One session on a GSM network issues several CDRs. As the subscriber moves from one routing area to another the CDRs must be correlated,” he says. “The CDRs arrive from the GSM and IP networks at different times and must be combined to identify one session and one customer.”

Time to Get the Sandbags?

Again and again, billing and mediation providers confronting revenue loss point fingers at someone else. Mediation companies point out that the network elements cannot send—or else lose—the necessary data. Billing providers attest that mediation platforms are guilty of sending unusable records. To compensate for the known leakage areas, both groups are adding functionality that will help them capture and identify the upcoming event records.

Singling out one area as the lone culprit for revenue loss is impossible. Although most target the network as the major offender, Amdocs’ Leon Malalel, director of product management and corporate marketing, says that attitude is shortsighted.

“Everyone points to the network because it is a big, black hole and few understand what is going on there,” he says. “Instead they should look at the overall picture. Leakage can come from 10,000 places; to target only the network causes problems.”

With this many potential leaks, it’s no wonder that billing for 2.5G services has been difficult. Free services, bundled offerings and flat-rate fees are so far the norm. With these billing schemes in place, operators will have a difficult time putting a price on the “value” of proposed 2.5G value-added services.

RA Programs: Locating Loose Change and Hard Cash
Now that service providers don’t have any money, everyone wants to help them save it—for a small fee, of course.

Revenue assurance, a current industry buzzword, means everything and yet nothing. Depending on who’s talking, revenue assurance includes containing fraud, locating lost data records, revising inefficient business processes and anything associated with saving service providers money.

All of the big five consulting firms and a few product companies have introduced revenue assurance programs to help carriers find revenue that is currently being left on the table.
The cost for these services depends on how comprehensive the project is, but it can range from $50,000 to $1.5 million, says Norbert Scholz, a BSS analyst at Gartner, who just completed an extensive report on revenue assurance competency standards.

Scholz describes the least expensive revenue assurance services as quick fixes with no process reengineering or recommendations for improving processes. These programs detect revenue loss at the network elements and mediation devices by looking for dropped calls, problems with rating systems and incorrectly rated calls. “These are more like a billing review,” he explains.

Amdocs’ Leon Malalel, director of product management and corporate marketing, claims that this type of network-focused revenue assurance project is hardly thorough. “Revenue assurance should be managed and discussed from an overall business perspective, not just from the technological side,” he says. “Revenue assurance should be a reassessment of the business.” Amdocs has identified 16 areas that must be addressed during a revenue assurance evaluation: sales and marketing, customer care, order management, provisioning and maintenance, acquisition and formatting, rating, error management, billing and invoicing, A/R and collections, financials, network infrastructure, contract management, intercarrier and partner settlements, commissions, fraud management and retail services.

Cap Gemini Ernst & Young also offers revenue assurance programs that review the entire landscape. Its program assesses marketing, sales and customer contact, order and provisioning, usage collection, usage assurance, billing, fraud management and receivables management.

These preventative and proactive measures do make for sound business. But are operators in highly competitive markets, like wireless, paying attention to them? These carriers must weigh time to market against prudent planning and establishing risk-averse environments. Are they even open to discussing these revenue assurance programs?

Gartner’s Scholz says that most operators are preoccupied with rolling out new services, finding the killer app and increasing subscribers. But some attitudes may be changing. “As penetration rates get higher and money gets tighter,” he says, “operators are having to look elsewhere to add revenue.”

Vibrant Solutions’ Yves Robinson, director of product management, is also seeing a shift in perspective. “Operators are being forced to deal with issues internally rather than expand services,” Robinson says. “They are being told to clean up the shop and increase the level of quality. For the first time revenue assurance managers are starting to find the necessary buy-in to introduce new processes and procedures.”
< BR>
Comments